Marine data collection for marine artificial intelligence systems

Abstract

A method comprising, by at least one processing unit, obtaining data collected by one or more sensors of at least one marine vessel, said data being representative of one or more situations encountered by the marine vessel during its voyage, prioritizing data according to at least one relevance criterion, wherein when the marine vessel is located in a zone in which at least one remote data communication link meets a criterion, transmitting at least some of the data using the at least one remote data communication link, wherein data are transmitted according to priority determined for the data, thereby facilitating transmission of relevant data for the purpose of training one or more machine learning algorithms (e.g. deep learning algorithms) providing output based on these data.

The invention claimed is: 
     
       1. A method, comprising:
 by at least one processing unit:
 obtaining data collected by one or more sensors of at least one marine vessel, said data being representative of one or more situations encountered by the marine vessel during a voyage thereof; 
 prioritizing the data according to at least one relevance criterion; and 
 when the marine vessel is located in a zone in which at least one remote data communication link meets a criterion, transmitting at least some of the data using the at least one remote data communication link, wherein the at least some of the data are transmitted according to priority determined for the data, 
 
 wherein prioritizing the data according to the at least one relevance criterion depends on a matching between: 
 one or more actual situations encountered by the marine vessel during the voyage thereof, and 
 one or more outputs provided by at least one machine learning algorithm, or by a system comprising said at least one machine learning algorithm, based at least on the data collected by said one or more sensors, the one or more outputs being representative of the one or more actual situations encountered by the marine vessel during the voyage thereof. 
 
     
     
       2. The method of  claim 1 , wherein the relevance criterion reflects at least one of:
 (i) a relevance of the data for training one or more given machine learning algorithms based on these data, the one or more given machine learning algorithms comprising the at least one machine learning algorithm, or being different from the at least one machine learning algorithm; or 
 (ii) a severity of a situation or an event encountered by the marine vessel during the voyage thereof. 
 
     
     
       3. The method of  claim 1 , wherein at least one of (i) or (ii) is met:
 (i) prioritizing the data according to the at least one relevance criterion depends on one or more actions taken by at least one of a crew or an auto-pilot of the marine vessel for controlling the marine vessel; or 
 (ii) prioritizing the data according to the at least one relevance criterion is based on at least a monitoring of data representative of at least one of a route or of inertial data of the marine vessel. 
 
     
     
       4. The method of  claim 1 , wherein when the marine vessel enters a zone in which remote data communication using at least one remote communication network meets a criterion, the method comprises transmitting at least some of the data over the remote communication network, wherein the at least some of the data are transmitted according to the priority determined for the data. 
     
     
       5. The method of  claim 1 , further comprising selecting a fraction of the obtained data for storing said fraction of obtained data before their transmission, wherein, for each period of time of a plurality of periods of time, a selection of the fraction of obtained data relative to the obtained data depends on priority determined for the obtained data associated with said period of time. 
     
     
       6. The method of  claim 1 , wherein the marine vessel embeds, or communicates with, the system comprising the at least one machine learning algorithm. 
     
     
       7. The method of  claim 6 , wherein the lower the matching between the one or more actual situations and the one or more outputs provided by the at least one machine learning algorithm, or by the system comprising the at least one machine learning algorithm, the higher the priority assigned to collected data representative of the one or more actual situations. 
     
     
       8. The method of  claim 1 , further comprising training at least one given machine learning algorithm based on the transmitted data, the at least one given machine learning algorithm corresponding to said at least one machine learning algorithm, or being different from the at least one machine learning algorithm. 
     
     
       9. A system, comprising:
 at least one processing unit configured to: 
 obtain data collected by one or more sensors of at least one marine vessel, said data being representative of one or more situations encountered by the marine vessel during a voyage thereof; 
 prioritize the data according to at least one relevance criterion; and 
 when the marine vessel is located in a zone in which at least one remote data communication link meets a criterion, trigger transmission of at least some of the data using the at least one remote data communication link, wherein the at least some of the data are transmitted according to priority determined for the data,
 wherein prioritizing the data according to the at least one relevance criterion depends on a matching between: 
 one or more actual situations encountered by the marine vessel during the voyage thereof, and 
 one or more outputs provided by at least one machine learning algorithm, or by a second system comprising at least one machine learning algorithm, based at least on the data collected by said one or more sensors, the one or more outputs being representative of the one or more actual situations encountered by the marine vessel during the voyage thereof. 
 
 
     
     
       10. The system of  claim 9 , wherein the relevance criterion reflects at least one of:
 (i) a relevance of the data for training one or more given machine learning algorithms based on these data, the one or more given machine learning algorithms comprising the at least one machine learning algorithm, or being different from the at least one machine learning algorithm; or 
 (ii) a severity of a situation or an event encountered by the marine vessel during the voyage thereof. 
 
     
     
       11. The system of  claim 9 , wherein prioritizing the data according to the at least one relevance criterion depends on one or more actions taken by at least one of a crew or an auto-pilot of the marine vessel for controlling the marine vessel. 
     
     
       12. The system of  claim 9 , wherein prioritizing the data according to the at least one relevance criterion is based on at least a monitoring of data representative of at least one of a route or of inertial data of the marine vessel. 
     
     
       13. The system of  claim 9 , wherein when the marine vessel enters a zone in which remote data communication using at least one remote communication network meets a criterion, the system is configured to trigger transmission of at least some of the data over the remote communication network, wherein the at least some of the data are transmitted according to the priority determined for the data. 
     
     
       14. The system of  claim 9 , configured to select a fraction of the obtained data for storing said fraction of obtained data before their transmission, wherein, for each period of time of a plurality of periods of time, selection of the fraction of obtained data relative to the obtained data depends on priority determined for the obtained data associated with said period of time. 
     
     
       15. The system of  claim 9 , wherein the system or the marine vessel;
 embeds, or communicates with, the second system comprising the at least one machine learning algorithm. 
 
     
     
       16. The system of  claim 15 , wherein the lower the matching between the one or more actual situations and the one or more outputs provided by the at least one machine learning algorithm, or by the second system, the higher the priority assigned to collected data representative of the one or more actual situations. 
     
     
       17. The system of  claim 15 , wherein prioritizing the data according to the least one relevance criterion for a specific sensor relative to data collected by other sensors depends on a matching between the one or more actual situations and the one or more outputs provided by the at least one machine learning algorithm, or by the second system for this specific sensor. 
     
     
       18. The system of  claim 9 , wherein relevance of data defined by the relevance criterion depends on one or more outputs of a plurality of different sensors of the marine vessel. 
     
     
       19. The system of  claim 9 , wherein at least one of (i), (ii), (iii), (iv) or (v) is met:
 (i) the at least one machine learning model, or the second system comprising the at least one machine learning model, is usable to detect objects around the marine vessel; 
 (ii) the at least one machine learning model, or the second system comprising the at least one machine learning model, is usable to assist navigation of the marine vessel; 
 (iii) the at least one machine learning model, or the second system comprising the at least one machine learning model, is usable to estimate distance to objects around the marine vessel; 
 (iv) the at least one machine learning model, or the second system comprising the at least one machine learning model, is usable to recognize objects around the marine vessel; 
 (v) the at least one machine learning model, or the second system comprising the at least one machine learning model, is usable to detect a potential collision between the marine vessel and one or more objects. 
 
     
     
       20. A non-transitory computer readable medium including instructions that, when executed by one or more processing circuitries, cause the one or more processing circuitries to perform a method comprising:
 obtaining data collected by one or more sensors of at least one marine vessel, said data being representative of one or more situations encountered by the marine vessel during a voyage thereof; 
 prioritizing data according to at least one relevance criterion; and 
 when the marine vessel is located in a zone in which at least one remote data communication link meets a criterion, transmitting at least some of the data using the at least one remote data communication link, wherein the at least some of the data are transmitted according to priority determined for the data, 
 wherein prioritizing the data according to the at least one relevance criterion depends on a matching between: 
 one or more actual situations encountered by the marine vessel during the voyage thereof, and 
 one or more outputs provided by at least one machine learning algorithm, or by a system comprising said at least one machine learning algorithm, based at least on the data collected by said one or more sensors, the one or more outputs being representative of the one or more actual situations encountered by the marine vessel during the voyage thereof.